From: Stephan Szabo <sszabo@megazone.bigpanda.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@postgresql.org
Subject: Re: [SQL] 7.4 - FK constraint performance
Date: Wed, 18 Feb 2004 08:03:36 -0800 (PST)
Message-ID: <20040218075937.T23729@megazone.bigpanda.com> (raw)
In-Reply-To: <20040215083314.B85170@megazone.bigpanda.com>
References: <20040213010341.95282.qmail@web60801.mail.yahoo.com>
<25808.1076634636@sss.pgh.pa.us>
<1076640825.10896.12.camel@jester>
<27443.1076646336@sss.pgh.pa.us>
<20040213065530.M19832@megazone.bigpanda.com>
<2455.1076686047@sss.pgh.pa.us>
<20040213073852.K20792@megazone.bigpanda.com>
<2662.1076687804@sss.pgh.pa.us>
<20040213080628.E21355@megazone.bigpanda.com>
<20040215083314.B85170@megazone.bigpanda.com>
On Sun, 15 Feb 2004, Stephan Szabo wrote:
> On Fri, 13 Feb 2004, Stephan Szabo wrote:
>
> >
> > On Fri, 13 Feb 2004, Tom Lane wrote:
> >
> > > Stephan Szabo <sszabo@megazone.bigpanda.com> writes:
> > > > On Fri, 13 Feb 2004, Tom Lane wrote:
> > > >> I was looking at that last night. It seems like we could add a LIMIT at
> > > >> least in some contexts. In the case at hand, we're just going to error
> > > >> out immediately if we find a matching row, and so there's no need for
> > > >> FOR UPDATE, is there?
> > >
> > > > I think there still is, because a not yet committed transaction could have
> > > > deleted them all in which case I think the correct behavior is to wait and
> > > > if that transaction commits allow the action and if it rolls back to
> > > > error.
> > >
> > > Good point. Okay, we can't put in a LIMIT. But we could still hack the
> > > planner to prefer a fast-start plan by passing an out-of-band tuple
> > > fraction, for those RI plans where it's appropriate. That would not
> > > affect correctness.
> >
> > Right, I can try to look through the stuff you pointed at in the previous
> > message over the weekend.
>
> It looks to me that we could make this available to SPI fairly simply by
> taking the current version of the following four routines: planner,
> pg_plan_query, _SPI_execute and SPI_prepare, renaming them and giving them
> a planning tuple fraction as a parameter, change references to the other
> routines to the new names and then making four new functions with the
> current names that call the renamed versions. In all the cases other than
> planner I think we can have the new version pass 0.0 and in the case of
> planner either 0.1 or 0.0 based on the isCursor parameter.
I did this, and changed the foreign keys to use it, but I haven't managed
to build a fk case where I could actually detect a change in the plan
chosen. Since the queries are only a simple scan on the one table I'm
wondering if it's basically just modifying both costs by the same value
which means there's no real effect at all.
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: pgsql-sql@postgresql.org
Cc: sszabo@megazone.bigpanda.com, tgl@sss.pgh.pa.us, pgsql-hackers@postgresql.org
Subject: Re: [SQL] 7.4 - FK constraint performance
In-Reply-To: <20040218075937.T23729@megazone.bigpanda.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox